nixos-rebuild with --target-host hangs in github action environment #262686
Comments
|
BTW, looking at your logs it seems I missed to preserve the bootloader updates inside the |
|
This works fine to me: @duament Can you pass the |
I can only reproduce it without the --tty flag. Here's the logs The second command without |
I believe it's not related to a systemd update since I can reproduce it by running the same command again without an actual system upgrade. |
Your configuration seems to be depending on See the reasoning here:
To workaround for now you can pass
Keep in mind that |
That $HOME related warning/error is from fish. It just complains my $HOME(/var/empty) is not writable. It will not exit with error. |
|
It seams |
|
Adding a I guess we should add a |
I guess it is fair: Especially because it seems |
|
I will test |
BTW, from the So adding both flags seems correct in this case. |
|
I tried to add the flag diff --git a/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh b/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
index 9e75db6d27b5..cb616d3cf709 100755
--- a/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
+++ b/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
@@ -663,6 +663,7 @@ if [[ "$action" = switch || "$action" = boot || "$action" = test || "$action" =
"--collect"
"--no-ask-password"
"--pty"
+ "--pipe"
"--quiet"
"--same-dir"
"--service-type=exec"But got stuck with some strange errors during tests: Not sure what is happening here, maybe @roberth can help? For now I think one workaround can be to wrap the call to |
|
BTW, I think we can substitute |
I think I found a way to fix the issue, opened a PR: #262985 |
I didn't work ( |
Investigating the test error: diff --git a/nixos/tests/nixos-rebuild-install-bootloader.nix b/nixos/tests/nixos-rebuild-install-bootloader.nix
index 3ade90ea24a..d1a01887f20 100644
--- a/nixos/tests/nixos-rebuild-install-bootloader.nix
+++ b/nixos/tests/nixos-rebuild-install-bootloader.nix
@@ -55,12 +55,12 @@ import ./make-test-python.nix ({ pkgs, ... }: {
"${configFile}",
"/etc/nixos/configuration.nix",
)
- machine.succeed("nixos-rebuild switch")
+ machine.succeed("stty -a 1>&2; nixos-rebuild switch; stty -a 1>&2; stty -echo")
# Need to run `nixos-rebuild` twice because the first run will install
# GRUB anyway
with subtest("Switch system again and install bootloader"):
- result = machine.succeed("nixos-rebuild switch --install-bootloader")
+ result = machine.succeed("stty -a 1>&2; nixos-rebuild switch --install-bootloader; stty -a 1>&2; stty -echo")
# install-grub2.pl messages
assert "updating GRUB 2 menu..." in result
assert "installing the GRUB 2 boot loader on /dev/vda..." in result
diff --git a/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh b/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
index 9e75db6d27b..cb616d3cf70 100755
--- a/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
+++ b/pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh
@@ -663,6 +663,7 @@ if [[ "$action" = switch || "$action" = boot || "$action" = test || "$action" =
"--collect"
"--no-ask-password"
"--pty"
+ "--pipe"
"--quiet"
"--same-dir"
"--service-type=exec"Found the tty settings being changed. Logs: Logs diff: --- old 2023-10-25 08:59:07.693220929 +0800
+++ new 2023-10-25 08:59:05.455250752 +0800
@@ -3,8 +3,8 @@
machine # eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R;
machine # werase = ^W; lnext = ^V; discard = ^O; min = 1; time = 0;
machine # -parenb -parodd -cmspar cs8 hupcl -cstopb cread -clocal -crtscts
-machine # -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl -ixon -ixoff
-machine # -iuclc -ixany -imaxbel -iutf8
-machine # -opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
-machine # -isig -icanon iexten -echo echoe echok -echonl -noflsh -xcase -tostop -echoprt
+machine # -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl -ixon -ixoff
+machine # -iuclc -ixany imaxbel iutf8
+machine # opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
+machine # isig icanon iexten echo echoe echok -echonl -noflsh -xcase -tostop echoprt
machine # echoctl echoke -flusho -extprocI'll try to restore the settings and see if the test could pass. |
|
I'm getting an identical hang using Packer to produce EC2 EBS images; with a Packer HCL fragment containing: I'm starting from the NixOS-23.05.555.52869451b83-* AMIs. I can confirm the |
|
Since upgrading to it works fine, but if I run then it hangs after this: until I ctrl-C it. Looking at and indeed (I assume the fact that I'm passing in a different nixpkgs with I have an effective workaround for my case (just running commands in |
Likely related to: NixOS/nixpkgs#262686
Likely related to: NixOS/nixpkgs#262686
Describe the bug
I've set up a github action to deploy nixos on remote hosts with command
But after #258571, it stops working.
The failed job https://github.com/duament/flakes/actions/runs/6598694622/job/17926924711 shows it hangs for about 2 hours and exit abnormally. On the target host I found
nixos-rebuild-switch-to-configuration.serviceexited successfully but thesystemd-runprocess hangs.I can also reproduce it locally with
systemd-run --user --same-dir nix run nixpkgs#nixos-rebuild -- --flake .#az --target-host deploy@az.rvf6.com --use-remote-sudo switch.Might be related to tty.
Steps To Reproduce
Steps to reproduce the behavior:
Expected behavior
Exit successfully
Screenshots
Additional context
Notify maintainers
@thiagokokada
Metadata
Please run
nix-shell -p nix-info --run "nix-info -m"and paste the result.The text was updated successfully, but these errors were encountered: